Sept ember Oct ober 2008

|

PCI Journal 54
Editors quick points
n This paper proposes provisions to extend the application of
current American Association of State Highway and Transporta-
tion Ofcials AASHTO LRFD Bridge Design Specications to
high-strength concrete (HSC) girders.
n The proposed design provisions are for predicting the ultimate
exural strength of prestressed concrete girders with concrete
compressive strengths up to 18 ksi (124 MPa) to include com-
posite action with normal-strength concrete (NSC) deck slabs.
n The experimental program investigated the failure modes of
three different types of compression zones: one with NSC only,
one with HSC only, and one with both NSC and HSC.
Behavior
and design of
high-strength
prestressed
concrete
girders
Wonchang Choi,
Sami Rizkalla,
Paul Zia,
and Amir Mirmiran
Although a number of state departments of transportation
have successfully used high-strength concrete (HSC) gird-
ers as part of Federal Highway Administration (FHWA)
demonstration projects, the current American Association
of State Highway and Transportation Officials AASHTO
LRFD Bridge Design Specifications
1
are limited to ap-
plications where concrete compressive strengths are 10 ksi
(69 MPa) or less.
Concrete with compressive strengths greater than 10 ksi
(69 MPa) is now available commercially as a result of im-
provements in concrete admixtures and the quality control
process in plants. Many researchers have shown that by us-
ing HSC, engineers are able to design bridges with longer
spans for a given girder cross section or reduce the number
of girders by increasing the girder spacing. Adelman and
Cousins
2
showed that increasing the concrete design com-
pressive strength from 6000 psi to 8000 psi (42 MPa to
55 MPa) resulted in an average 10% increase in span capa-
bility for prestressed girders used in routine bridge design.
Due to these advantages, it is likely that the use of HSC in
the design of prestressed girders will continue to increase.
However, the uncertainty regarding the applicability of de-
sign provisions causes reluctance on the part of designers
to use HSC for highway bridge construction.
3
Therefore, a
need exists for reassessment of the material properties and
the design provisions for the analysis of HSC girders. This
need to expand the applicability of the AASHTO LRFD
specifications to HSC has been addressed by a series of
projects under the direction of the National Cooperative
Highway Research Program (NCHRP). The goal of one of
these projects, NCHRP project 12-64, was to expand the
use of the AASHTO LRFD specifications to reinforced
54
55 PCI Journal

|

Sept ember Oct ober 2008
the girder designs required sixteen to twenty in.
(13 mm), grade 270 (1860 MPa), 7-wire strands. Each
strand was tensioned to 75% of its ultimate strength, or 31
kip (138 kN). All strands were straight and fully bonded
over their entire lengths. Figure 1 shows strand configura-
tions for the three design concrete compressive strengths.
After concrete was placed in the girders for each design
concrete compressive strength, the concrete was moist-
cured until it reached the required releasing strength. After
release of the prestressing strands, the concrete girders
were air-cured in the prestressing plant until they were
delivered to the laboratory.
For each HSC test girder, fifteen 4 in. 8 in. (100 mm
200 mm) cylinders and nine 6 in. 6 in. 20 in.
(150 mm 150 mm 500 mm) prisms were made for
each casting to determine the HSCs elastic modulus and
modulus of rupture, respectively. The cylinder and prism
specimens were cured with the test girders and under the
same conditions.
After the girders were delivered to the laboratory, a 5-ft-
wide (1.5 m) deck slab was cast on one of the three gird-
ers, a 1-ft-wide (0.3 m) deck slab was cast on the second
girder, and the third girder was left without a deck slab
for each of the target concrete compressive strengths.
Figure 1 shows the final cross sections for each specimen
group.
A local ready-mix concrete producer supplied the
concrete used to cast the 8-in.-thick (200 mm) deck
slabs. The average 28-day compressive strengths of
the concrete used for the 5-ft- and 1-ft-wide (1.5 m and
0.3 m) deck slabs were 4.1 ksi and 5.6 ksi (28 MPa and
38 MPa), respectively.
and prestressed concrete members with 18 ksi (124 MPa)
compressive strength in flexure and compression.
As a part of NCHRP project 12-64, research was conducted
to examine the validity of the current analytical methods
used to determine the flexural behavior of typical AASHTO
Type II prestressed HSC girders with and without a cast-
in-place normal-strength concrete (NSC) deck. The ex-
perimental program validated the stress block parameters
currently used to determine the flexural resistance of flanged
sections made with HSC. It also addressed the cases where
the compression zone was composed of NSC deck and HSC
girder in composite action. Test results were used to evaluate
the AASHTO LRFD specifications equations to predict the
elastic modulus and modulus of rupture for HSC.
Experimental program
Nine 40-ft-long (12 m) AASHTO Type II prestressed HSC
girders were designed and tested to evaluate their flex-
ural responses. They were tested under a static load using
four-point bending. The concretes used for the girders were
designed to achieve target compressive strengths of 10
ksi, 14 ksi, and 18 ksi (69 MPa, 97 MPa, and 124 MPa).
Table 1 shows the mixture proportions of the concretes.
All girders were designed based on the AASHTO LRFD
specifications. However, several design details were modi-
fied to prevent premature failure in shear or bond slippage
prior to flexural failure. Shear reinforcement consisted of
no. 4 (13M) stirrups at a spacing of 3 in. (75 mm) near the
end blocks and 6 in. (150 mm) along the rest of the girder.
More detailed information about the test girders can be
found in Choi.
4
Standard Concrete Products in Savannah, Ga., produced
the girder specimens. The three concrete strengths used for
Table 1. Mixture properties for prestressed, AASHTO Type II high-strength concrete girders
Target compressive strength 10 ksi 14 ksi 18 ksi
Cement, lb/yd
3
670.0 703.0 890.0
Fly ash, lb/yd
3
150.0 192.0 180.0
Microsilica, lb/yd
3
50.0 75.0 75.0
No. 67 granite, lb/yd
3
1727.0 1700.0 1700.0
Sand (river), lb/yd
3
1100.0 1098.0 917.0
Water, lb/yd
3
280.0 250.0 265.0
Recover hydration stabilizer, oz/yd
3
26.0 50.0 50.0
Waterreducing admixture, oz/yd
3
98.0 125.0 135.0
Watercementitious materials ratio 0.32 0.26 0.23
Source explanation for no. 67: American Society for Testing and Materials (ASTM). 2007. Standard Specication for Concrete Aggregate. ASTM C33-07.
West Conshohocken, PA: ASTM.
Note: 1 ksi = 6.895 MPa; 1 lb/yd
3
= 0.5933 kg/m
3
; 1 oz/yd
3
= 38.7 mL/m
3
.
Sept ember Oct ober 2008

|

PCI Journal 56
Figure 1. These diagrams illustrate the strand congurations for 18 ksi, 14 ksi, and 10 ksi design concrete compressive strengths and the cross sections for an 18 ksi
high-strength concrete girder with different deck-slab congurations. Note: HSC = high-strength concrete. 1 in. = 25.4 mm; 1 ft = 0.3048 m; 1 ksi = 6.895 MPa.


36 in.
8 in.
18 in.
6 in.
60 in.

36 in.
12 in.
18 in.
8 i n.
6 in.

36 in.
12 in.
18 in.
6 in.
18 ksi concrete girder 14 ksi concrete girder 10 ksi concrete girder
18 ksi HSC girder with a 5-ft-wide deck slab
18 ksi HSC girder with a 1-ft-wide deck slab
18 ksi HSC girder without a deck slab
57 PCI Journal

|

Sept ember Oct ober 2008
statistical analysis are presented in previous studies.
4,8
E
c
= 310,000K
1
w
c
2.5

f
c
'
0.33
(1)
where
K
1
= correction factor to account for aggregate source
w
c
= density of concrete

f
c
'
= specified design compressive strength of concrete
The correction factor K
1
is typically assumed to be 1.0
unless determined by physical testing and as approved by
the authority of jurisdiction. The results indicated that the
AASHTO LRFD specifications overestimate the elastic
modulus determined from HSC cylinder tests, while the
proposed equation provides a closer prediction.
Figure 5 shows the average moduli of rupture f
r
obtained
from the two control specimens for each of the nine girder
specimens. Predictions of the modulus of rupture using the
two expressions given by the AASHTO LRFD speci-
fications are also plotted in the same figure. One of the
AASHTO expressions is used for computing cracking mo-
ment under service limit load combination (f
r
= 0.24 f
c
'

[in ksi]), while the other is used for determining minimum
reinforcement (f
r
= 0.37 f
c
'
[in ksi]). Test results con-
firmed that the current equations of the AASHTO LRFD
specifications overestimated the modulus of rupture for
HSC. Equation (2), which is from ACI 318-05, is therefore
recommended to estimate the modulus of rupture for HSC
up to 18 ksi (124 MPa):
f
r
= 0.19 f
c
'
(in ksi) (2)
Five or six characters identify each girder specimen. The
first two numbers represent the design concrete compres-
sive strength of the girders, followed by the uppercase
letters PS, which stands for prestressed concrete girder.
The final one or two characters following the hyphen
show the dimensions of the deck slab, using a number to
represent the deck width, with either an uppercase letter S
to represent slab or an uppercase letter N to represent no
slab. For example, 10PS-5S represents a prestressed girder
made with 10 ksi design concrete compressive strength and
a 5-ft-wide (1.5 m) deck slab.
Load cells monitored the prestressing force in each girder
from the start of fabrication to the time immediately
before the transfer of prestress. The prestressing force was
checked against the elongation of selected strands at the
time of jacking. Prior to placing the concrete, two strain
gauges were welded (using low voltage) to two strands
at the bottom row of each girder to measure the strain
changes in the prestressing strands due to elastic shorten-
ing, prestress losses, and strains in prestressing strands
during load tests. Figure 2 shows an installed strain gauge.
Figure 3 shows a schematic view of the test setup and a
photo of the typical instrumentation layout. Potentiometers
measured the deflections at midspan, loading points, and
quarter points along the girder and at the supports. Load
was applied in displacement control at a rate of
0.1 in./min (2.5 mm/min) in order to observe crack initia-
tion in the girder. The loading rate was increased to 0.25
in./min (6.3 mm/min) after the prestressing strands yielded.
Visual inspection and mapping of the cracks were per-
formed throughout the tests. A high-speed data acquisition
system was used to record data from the potentiometers
and strain gauges. Tests were terminated after crushing of
concrete occurred in the constant-moment region.
Material properties
and early-age behavior
Table 2 lists the measured material properties of the
concrete on the test day for each girder and deck slab. All
tests for material properties conformed with ASTM speci-
fications (ASTM C39,
5
ASTM C469,
6
and ASTM C78
7
).
Except for concrete used for girder 18PS-1S, all specimens
achieved their design compressive strengths.
Table 2 also lists the average elastic moduli obtained from
the three control cylinders for each of the nine prestressed
AASHTO girders tested for this project. Figure 4 shows
the predicted values using the AASHTO LRFD specifica-
tions as well as the predictions according to Eq. (1) using
149 lb/ft
3
(2387 kg/m
3
) as the measured density of the con-
crete. Equation (1) is the proposed formula
4,8
to determine
the elastic modulus E
c
for HSC with strengths ranging from
10 ksi to 18 ksi (69 MPa to 124MPa). It was obtained by a
statistical analysis of over 4000 test results. Details of the
Figure 2. The photo shows an example of the strain gauges that were welded
to two of the bottom strands in each girder to measure the strain changes in the
prestressing strands due to elastic shortening, prestress losses, and strains in
prestressing strands during load tests.
Sept ember Oct ober 2008

|

PCI Journal 58
where
E
p
= modulus of elasticity of strand
= strand end slippage
f
pi
= initial prestress of the strand just before detensioning
Transfer length
To determine the transfer lengths, the end slippages of six
preselected strands were measured using a tape measure-
ment before and after prestress transfer. Equation (3)
9
was
then used to determine the transfer lengths l
t
.
l
t
=
2E
f
p
pi

(3)
Figure 3. Shown are a schematic view of the test setup and an instrumentation layout and a photo of the test setup for a girder with a 5-ft-wide deck slab.
Note: CFL = Constructed Facilities Laboratory. 1 ft = 0.3048 m.
Potentiometers

Potentiometers Potentiometers

40 ft
6 ft

10 ft
CFL floor
59 PCI Journal

|

Sept ember Oct ober 2008
Girder tests
Load-deflection response
and failure mode
Figure 6 presents the load-deflection responses at the
midspans of the three girders with 5-ft-wide (1.5 m) deck
slabs. The figure shows that the initial flexural stiffnesses
of all three composite girders were practically the same
prior to cracking. Also, the flexural stiffnesses were not
affected by the compressive strength of concrete because
there were only small differences in the elastic moduli
of the three different concretes. In addition, Fig. 6 shows
the predicted load-deflection responses using a section
analysis program, RESPONSE.
10
The predicted responses
agree with the measured responses for the flexural stiffness
initially and at ultimate load, while the measured deflec-
tion after yielding of prestressing strands is slightly less
than the predicted deflection, possibly due to the variation
of prestrain for input data. Figure 6 shows that the failure
of test specimens occurred gradually due to crushing of
concrete within the NSC deck slab.
Figure 7 shows the measured and predicted load-
deflection responses of the three composite girders with
1-ft-wide (0.3 m) deck slabs. A similar behavior among the
three girders was observed prior to the initiation of cracks.
The measured responses reflected a small drop in load-
Table 3 lists the calculated transfer lengths of the girder
specimens. The range of the measured transfer lengths
varied from 21 in. to 34 in. (525 mm to 850 mm). Based
on these data, it appears that the predicted value of 30 in.
(750 mm) for -in.-diameter (13 mm) strand by AASHTO
LRFD specifications section 5.11.4.1 is reasonable for
concrete compressive strengths up to 18 ksi (124 MPa).
Elastic shortening
Table 4 compares the measured and calculated losses due
to elastic shortening at the bottom level of prestressing
strands. The calculated values were based on the AASHTO
LRFD specifications Eq. (5.9.5.2.3a-1) with the elastic
modulus specified by the current AASHTO LRFD specifi-
cations, as well as the proposed Eq. (1).
The table indicates that the average loss due to elastic
shortening at the bottom level of strands was 7.7%, which
is close to the predicted values using the current AASHTO
LRFD specifications as well as Eq. (1).
Table 2. Material properties of test specimens
Specimen Specimen type Age, days f
c(test)
, ksi E, ksi f
r
, ksi
10PS-5S
Girder 120 11.49 5360 0.768
Deck 29 3.78 2690
14PS-5S
Girder 143 16.16 5560 0.711
Deck 43 5.34 3300
18PS-5S
Girder 175 18.06 5970 0.872
Deck 67 3.99 2660
10PS-1S
Girder 189 13.19 5630 0.820
Deck 77 5.04 2770
14PS-1S
Girder 184 15.53 5440 0.751
Deck 70 5.04 2770
18PS-1S
Girder 199 14.49 5150 0.680
Deck 84 5.04 2770
10PS-N Girder 222 11.81 5540 0.820
14PS-N Girder 228 15.66 5330 0.717
18PS-N Girder 232 18.11 6020 0.706
Note: 1 ksi = 6.895 MPa.
Sept ember Oct ober 2008

|

PCI Journal 60
Figure 5. This graph plots the concrete compressive strength versus the modulus of rupture to compare the average moduli of rupture f
r
obtained from the two control
specimens for each of the nine girder specimens with the predictions from the two expressions given by the American Association of State Highway and Transportation
Ofcials AASHTO LRFD Bridge Design Specications. Note: Equations are in English units.1 ksi = 6.895 MPa.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
2
4 6 8 10 12 14 16 18 20
Concrete compressive strength f'
c, ksi
M
o
d
u
l
u
s

o
f

r
u
p
t
u
r
e

f
r
,

k
s
i
0
2
4
6
8
10
12
28 48 68 88 108 128
Concrete compressive strength f' c
, MPa
M
o
d
u
l
u
s

o
f

r
u
p
t
u
r
e

f
r
,

M
P
a
This research
ACl 318-02 (strength)
AASHTO LRFD
(for cracking moment in deflection calculations)
f
r
= 0.24 f'
c
(ksi)
AASHTO LRFD
(for cracking moment in minimum reinforcement calculations)
f
r = 0.37 f'
c
(ksi)
Proposed equation
f
r = 0.19 f'
c
(ksi)
Figure 4. This graph plots the concrete compressive strength versus the elastic modulus to compare the research results with the predicted values using the American
Association of State Highway and Transportation Ofcials AASHTO LRFD Bridge Design Specications and the proposed equation E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
.
Note: Equations are in English units. 1 kip-ft = 1.356 kN-m; 1 ksi = 6.895 MPa.
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
6 8 10 12 14 16 18 20
Concrete compressive strength f'
c
, ksi
E
l
a
s
t
i
c

m
o
d
u
l
u
s


E
c
,

k
s
i






0
10,000
20,000
30,000
40,000
50,000
60,000
41 56 71 86 101 116 131
E
l
a
s
t
i
c

m
o
d
u
l
u
s

E
c
,

M
P
a








This research
Proposed equation
E

= 310,000(wc)
2.5
(f'
c
)
0.33
AASHTO LRFD
E
c
= 33,000(w
c
)
1.5
f'c
Concrete compressive strength f'
c
, MPa
AASHTO LRFD
Proposed equation
61 PCI Journal

|

Sept ember Oct ober 2008
carrying capacity near failure due to complete crushing of
the NSC deck slab followed by crushing of a portion of
the HSC girder flange. Failure of test specimens occurred
suddenly after crushing of the deck slab followed by crush-
ing of the top flange of the HSC girder. Figure 7 shows the
buckling of the longitudinal reinforcement and prestressing
strand in the compression zone.
Figure 8 shows the measured and predicted load-
deflection responses of the three girders without a deck
slab. The three girders without a deck slab exhibited simi-
lar behavior to that of the composite HSC girders except
that the failure mode was more brittle. For the girders
without a deck slab, Fig. 8 shows that the failure occurred
suddenly, followed by the buckling of prestressing strand
in the compression zone. In two of the three casesgirder
without a deck slab and girder with a 1-ft-wide (0.3 m)
deck slabthe sudden crushing of the compression zone
also led to immediate crushing of concrete in the web.
Cracking moment
Table 5 compares the measured cracking moments M
cr
of
the nine girders with the calculated values using Eq. (4).
M
cr
= S
bc
(f
r
+ f
ce
- f
d/nc
) (4)
where
S
bc
= composite section modulus
f
ce
= compressive stress due to effective prestress only at
the bottom fibers
f
d/nc
= stress due to non-composite dead loads at the same
load level
When predicting the cracking moment, two different values
were used for f
r
: that specified by the current AASHTO
LRFD specifications article 5.4.2.6 and the one recommend-
ed in

Eq. (2). The predicted cracking moment depends on
the modulus of rupture. For all girder specimens, the results
continued on page 65
Table 3. End slippage and transfer lengths of test specimens
Specimen , in. l
t
, in.
18PS-1S 0.10 29.0
18PS-5S 0.08 23.0
18PS-N 0.04 12.0
Average 0.07 21.3
14PS-1S 0.08 22.0
14PS-5S 0.10 29.0
14PS-N 0.13 36.0
Average 0.10 29.0
10PS-1S 0.10 30.0
10PS-5S 0.20 58.0
10PS-N 0.05 15.0
Average 0.12 34.3
Note: 1 in. = 25.4 mm.
Table 4. Elastic shortening loss at the bottom-level strands of test specimens
Specimen Initial strain,
Average measured losses
Calculated losses from elastic modulus, %
AASHTO LRFD Bridge
Design Specications
Proposed equation
E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
%
18PS-1S 6282 613 9.8 7.9 7.9
18PS-5S 6149 329 5.3 7.8 7.9
18PS-N 5786 401 6.9 7.6 7.7
14PS-1S 6189 495 8.0 7.7 7.5
14PS-5S 6282 448 7.1 8.2 7.9
14PS-N 5579 482 8.6 8.2 7.9
10PS-1S 6373 522 8.2 6.8 6.6
10PS-5S 6333 532 8.4 6.8 6.6
10PS-N 6477 472 7.3 6.7 6.6
Average 6161 477 7.7 7.5 7.4
Sept ember Oct ober 2008

|

PCI Journal 62
Figure 6. The graph plots the load-deection responses, and the photo shows a typical failure mode of the three girders with 5-ft-wide deck slabs. Note: 1 ft = 0.3048 m.
0
50
100
150
200
250
300
0 2 4 6 8 10 12 14
Deflection at midspan, in.
L
o
a
d
,

k
i
p
0
200
400
600
800
1000
1200
0 50 100 150 200 250 300 350
Deflection at midspan, mm
L
o
a
d
,

k
N
10PS-5S
14PS-5S
18PS-5S
Measured
Predicted
63 PCI Journal

|

Sept ember Oct ober 2008
Figure 7. The graph plots the load-deection responses, and the photo shows a typical failure mode for the three girders with 1-ft-wide deck slabs. Note: 1 ft = 0.3048 m.
0
50
100
150
200
250
300
0 2 4 6 8 10 12 14
Deflection at midspan, in.
L
o
a
d
,

k
i
p
0
200
400
600
800
1000
1200
0 50 100 150 200 250 300 350
Deflection at midspan, mm
L
o
a
d
,

k
N
10PS-1S
14PS-1S
18PS-1S
Measured
Predicted
Sept ember Oct ober 2008

|

PCI Journal 64
Figure 8. The graph plots the load-deection responses, and the photo shows a typical failure mode for the three girders without deck slabs.
0
50
100
150
200
250
300
0 2 4 6 8 10 12 14
Deflection at midspan, in.
L
o
a
d
,

k
i
p
0
200
400
600
800
1000
1200
0 50 100 150 200 250 300 350
Deflection at midspan, mm
L
o
a
d
,

k
N
10PS-N
14PS-N
18PS-N
Measured
Predicted
65 PCI Journal

|

Sept ember Oct ober 2008
the short-term prestress losses occurring from the time of
release to the time of testing and do not include the long-
term prestress losses of the nine girder specimens.
Flexural strength
The flexural strengths of all girder specimens were
calculated using three different approaches. In the first ap-
proach, the AASHTO LRFD specifications Eq. (5.7.3.2.2-
1) was used with the current values of
1
and
1
in the
specification. In the second approach, Eq. (5) and (6) were
used to determine
1
and
1
as the new recommended
relationships.
8,11,12

1
=

0.85 for f
c
'
10 ksi
0.85 0.02 f
c
'
10
( )
0.75 for f
c
'
10 ksi
(5)
continued from page 61
in Table 5 indicate that the predicted cracking moment using
the proposed modulus of rupture produced conservative
results. The results suggest that the recommended Eq. (2) for
the modulus of rupture is more appropriate to determine the
cracking moment of prestressed HSC girders.
After initial cracking was observed, each test girder was
unloaded. On the girders second loading, the moment that
caused the crack to reopen was recorded. Based on these
two moment measurements and the rupture modulus, the
loss of prestress at the time of test was calculated.
8
The
calculated prestress loss varied from 7.3% to 13.9% for the
nine test girders, with an average of 11%. This loss com-
pares with an average of 15.1% based on AASHTO LRFD
specifications and an average of 14.9% based on Eq. (2)
for the rupture modulus.
8
These loss values represent only
Table 5. Summary of measured and predicted cracking moments of test specimens
Specimen
Measured
cracking moment
Predicted cracking moment
AASHTO LRFD Bridge Design Specications Proposed equation E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
kip-ft kip-ft
Measured
Predicted
kip-ft
Measured
Predicted
10PS-5S 1097 1123 0.98 1061 1.03
14PS-5S 1267 1314 0.96 1244 1.02
18PS-5S 1377 1436 0.96 1373 1.00
10PS-1S 935 974 0.96 922 1.01
14PS-1S 1054 1084 0.97 1034 1.02
18PS-1S 1131 1183 0.96 1130 1.00
10PS-N 799 751 1.06 708 1.13
14PS-N 867 843 1.03 796 1.09
18PS-N 918 964 0.95 908 1.01
Note: 1 kip-ft = 1.356 kN-m.
Table 6. Flexural strength of girders with a 5-ft-wide deck
Specimen
Measured
moment
Flexural strength
AASHTO LRFD Bridge Design
Specications
Proposed equation
E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
Strain compatibility
kip-ft kip-ft
Measured
Predicted
kip-ft
Measured
Predicted
kip-ft
Measured
Predicted
10PS-5S 2123 1904 1.12 1904 1.12 1977 1.07
14PS-5S 2349 2181 1.08 2181 1.08 2246 1.05
18PS-5S 2543 2344 1.08 2344 1.08 2445 1.04
Note: 1 ft = 0.3048 m; 1 kip-ft = 1.356 kN-m.
Sept ember Oct ober 2008

|

PCI Journal 66
LRFD specifications can be used to predict the flexural
strength when the compression zone is within the NSC
deck slab.
For the composite girders with the 1-ft-wide (0.3 m) deck
slabs, the current AASHTO LRFD specifications do not
provide clear recommendations on how to determine the
flexural strength of a section when the compression zone
includes two different concrete compressive strengths.
Because the neutral axis was located below the deck, the
compression zone required two different concrete stress-
strain distributions. However, for simplicity, the stress
distribution in the compression zone may be assumed
conservatively using the stress-strain relationship of NSC.
Therefore, the equivalent rectangular stress block can be
determined with the recommended method.
The computed flexural strengths using the recommended
method were about 12% to 14% less than the measured
flexural resistance. These results in Table 7 indicated that
this method can be used to safely determine the nomi-
nal flexural strength M
n
. The predicted nominal flexural
strength based on the strain compatibility with the mea-
sured material properties showed more accurate results
within a 1% difference of the measured flexural strength.
Table 8 gives the measured and predicted ultimate flexural
strengths using the proposed
1
and
1
and the strain
compatibility for the girders without a deck slab. This table

1
=

0.85 for f
c
'
4ksi
0.85 0.05 f
c
'
4
( )
0.65 for f
c
'
4 ksi
(6)
where

f
c
'
is in ksi

1
= stress-block parameter

1
= stress-block parameter
The third approach was based on strain compatibility, force
equilibrium, and the actual stress-strain relationship of the
concrete obtained from tests of control cylinders.
For the composite girders with the 5-ft-wide (1.5 m) deck
slabs, the flexural strength depended on whether the neu-
tral axis was located in the flange or in the girder. Because
the neutral-axis depth c was located in the deck concrete,
the composite girder behaved as a rectangular section. The
stress-block parameters for computing the flexural strength
of the composite HSC girders could be determined using
the current AASHTO LRFD specifications. Table 6 shows
the comparisons between the measured and predicted val-
ues of flexural strength of the three girders with 5-ft-wide
deck slabs using the three approaches mentioned previ-
ously. The comparisons indicate that the current AASHTO
Table 7. Flexural strengths of girders with a 1-ft-wide deck
Specimen
Measured moment
Flexural strength
Proposed equation E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
Strain compatibility
kip-ft kip-ft
Measured
Predicted
kip-ft
Measured
Predicted
10PS-1S 1752 1558 1.12 1735 1.01
14PS-1S 1941 1706 1.14 1928 1.01
18PS-1S 2083 1830 1.14 2107 0.99
Note: 1 ft = 0.3048 m; 1 kip-ft = 1.356 kN-m.
Table 8. Flexural strengths of girders without deck slab
Specimen
Measured moment
Flexural strength
Proposed equation E
c
= 310,000K
1
w
c
2.5
f
'
c
0.33
Strain compatibility
kip-ft kip-ft
Measured
Predicted
kip-ft
Measured
Predicted
10PS-N 1465 1324 1.11 1433 1.02
14PS-N 1688 1519 1.11 1623 1.04
18PS-N 1808 1692 1.07 1813 1.00
Note: 1 kip-ft = 1.356 kN-m.
67 PCI Journal

|

Sept ember Oct ober 2008
References
1. American Association of State Highway and Trans-
portation Officials (AASHTO). 2004. AASHTO
LRFD Bridge Design Specifications. 3rd ed. Wash-
ington, DC: AASHTO.
2. Adelman, D., and T. E. Cousins. 1990. Evaluation of
the Use of High Strength Concrete Bridge Girders in
Louisiana. PCI Journal, V. 35, No. 5 (September
October): pp. 7078.
3. Roller, J. J., B. T. Martin, H. G. Russell, and R.
N. Bruce. 1993. Performance of Prestressed High
Strength Concrete Bridge Girders. PCI Journal, V.
38, No. 3 (MayJune): pp. 3445.
4. Choi, W. C. 2006. Flexural Behavior of Prestressed
Girder with High Strength Concrete. PhD thesis.
Department of Civil, Construction, and Environmen-
tal Engineering, North Carolina State University,
Raleigh, NC.
5. American Society for Testing and Materials (ASTM).
2005. Standard Test Method for Compressive
Strength of Cylindrical Concrete Specimens. ASTM
C39/C39M-05e1. West Conshohocken, PA: ASTM.
6. ASTM. 2002. Standard Test Method for Static Modu-
lus of Elasticity and Poissons Ratio of Concrete in
Compression. ASTM C469-02e1. West Conshohock-
en, PA: ASTM.
7. ASTM. 2008. Standard Test Method for Flexural
Strength of Concrete (Using Simple Beam with Third-
Point Loading). ASTM C78-08. West Conshohocken,
PA: ASTM.
8. Rizkalla, S., A. Mirmiran, P. Zia, H. Russell, and R.
Mast. 2007. Application of the LRFD Bridge Design
Specifications to High-Strength Structural Concrete:
Flexure and Compression Provisions. NCHRP report
595. Washington, DC: Transportation Research
Board, the National Academies.
9. Oh, B. H., and E. S. Kim. 2000. Realistic Evaluation
of Transfer Lengths in Pretensioned Prestressed Con-
crete Members. ACI Structural Journal, V. 97, No. 6
(NovemberDecember): pp. 821830.
10. Bentz, E. C. 2000. Sectional Analysis of Reinforced
Concrete Members. PhD thesis. Department of Civil
Engineering, University of Toronto, Toronto, ON.
11. Mertol, H. C. 2006. Characteristics of High Strength
Concrete for Combined Flexure and Axial Compres-
sion Members. PhD thesis. Department of Civil,
indicates that it is satisfactory to use the proposed parame-
ters
1
and
1
to predict the flexural strength of prestressed
girders with concrete strengths up to 18 ksi (124 MPa).
Conclusion
The flexural behaviors of prestressed HSC girders with and
without deck slabs were investigated, including the material
properties and their early-age behaviors. Based on the ex-
perimental results, the following conclusions were drawn:
The current AASHTO LRFD specifications equation
for the elastic modulus of concrete may overestimate
measured values. Based on the results from this study,
the recommended equation provides better agreement
with the measured values for HSC girders with con-
crete compressive strengths up to 18 ksi (124 MPa).
Based on the findings of this study, Eq. (2), which is
the ACI 318-05 calculation for the modulus of rupture,
provides a better estimate of the cracking moment for
prestressed HSC girders with concrete compressive
strengths up to 18 ksi (124 MPa) than the equation
provided by the AASHTO LRFD specifications.
The transfer-length equation of the AASHTO LRFD
specifications provides a reasonable estimate for
prestressed HSC girders with concrete compressive
strengths up to 18 ksi (124 MPa).
For a composite HSC girder with NSC deck slab, if
the compression zone occurs in both HSC and NSC, a
conservative estimate of the nominal flexural strength
can be determined based on the concrete compressive
strength of the NSC deck with reasonable accuracy.
When the compression zone occurs only in HSC, the
nominal flexural strength can be determined using the
AASHTO LRFD specifications with the recommended
values of
1
and
1
for HSC girders with concrete
compressive strengths up to 18 ksi (124 MPa).
Acknowledgments
The authors acknowledge the support of NCHRP project
12-64 and the senior program officer, David Beal. They
are also grateful for the contributions of Henry Russell
of Henry Russell Inc. and Robert Mast of Berger/ABAM
Engineers Inc., both of whom served as consultants for the
project. The cooperation of Standard Concrete Products
in Savannah, Ga., and the personnel of the Constructed
Facilities Laboratory are greatly appreciated. The authors
are responsible exclusively for the findings and opinions
expressed in this paper.
Sept ember Oct ober 2008

|

PCI Journal 68
Construction, and Environmental Engineering, North
Carolina State University, Raleigh, NC.
12. Mertol, H. C., S. Rizkalla, P. Zia, and A. Mirmiran.
2008. Characteristics of Compressive Stress Distri-
bution in High-Strength Concrete. ACI Structural
Journal (under review).
Notation
c = depth at neutral axis
E
c
= modulus of elasticity of concrete
E
p
= modulus of elasticity of strand

f
c
'
= specified compressive strength of concrete
f
ce
= compressive stress due to effective prestress only at
the bottom fibers
f
d/nc
= stress due to non-composite dead loads at the same
load level
f
pi
= initial prestress of the strand just before detensioning
f
r
= modulus of rupture
K
1
= the correction factor to account for aggregate source
l
t
= transfer length
M
cr
= cracking moment
M
n
= nominal flexural strength
S
bc
= composite section modulus
w
c
= density of concrete

1
= stress-block parameter

1
= stress-block parameter
= strand end slippage
69 PCI Journal

|

Sept ember Oct ober 2008
About the authors
Wonchang Choi, PhD, is an
adjunct assistant professor of
Civil, Construction, and
Environmental Engineering at
North Carolina State University
in Raleigh, N.C.
Sami Rizkalla, PhD, P.Eng., is a
Distinguished Professor of Civil,
Construction, and Environmental
Engineering and director of the
Constructed Facilities Labora-
tory at North Carolina State
University.
Paul Zia, PhD, P.E., FPCI, is a
Distinguished University Profes-
sor Emeritus at North Carolina
State University.
Amir Mirmiran, PhD, P.E., is a
professor and interim dean for
the College of Engineering and
Computing at Florida Interna-
tional University in Miami, Fla.
Synopsis
This paper proposes provisions to extend the current
American Association of State Highway and Trans-
portation Officials AASHTO LRFD Bridge Design
Specifications to include prediction of the ultimate
flexural strength of prestressed concrete girders
with concrete compressive strengths up to 18 ksi
(124 MPa). The proposed design provisions include
composite action of a high-strength concrete (HSC)
girder with normal-strength concrete (NSC) deck
slab.
Nine 40-ft-long (12 m) AASHTO Type II HSC gird-
ers were tested with and without cast-in-place NSC
decks of differing widths to achieve various possible
modes of failure. The concrete used for the girder
was designed for three target compressive strengths
of 10 ksi, 14 ksi, and 18 ksi (69 MPa, 97 MPa, and
124 MPa).
The experimental program investigated failure modes
of three different types of compression zones: one
with NSC only, one with HSC only, and one with
both NSC and HSC. All girders were tested to failure
under static loading to study the different limit-state
behaviors, including prestress losses, initiation of
cracking, yielding, and final failure mode.
Keywords
Bridge girder, cracking strength, elastic shortening,
flexural strength, high-strength concrete, prestress
loss, transfer length.
Review policy
This paper was reviewed in accordance with the
Precast/Prestressed Concrete Institutes peer-review
process.
Reader comments
Please address any reader comments to PCI Journal
editor-in-chief Emily Lorenz at elorenz@pci.org
or Precast/Prestressed Concrete Institute, c/o PCI
Journal, 209 W. Jackson Blvd., Suite 500, Chicago,
IL 60606. J